African Americans represent about 10% of the population in the US, but are under-represented in biomarker- related aging studies such as the Alzheimer's Disease Neuro-imaging Initiative (ADNI) and World Wide ADNI. Epidemiologic studies show that, compared to non-Hispanic white (NHW) Americans, African Americans (AA) are more likely to develop mild cognitive impairment (MCI) and Alzheimer's disease (AD), have different genetic risks of developing AD, and experience different rates of cognitive decline after cognitive symptoms develop. All these point to the existence of an MCI/AD endophenotype for AA, although few of these epidemiological studies involve modern chemical or imaging biomarkers associated with AD pathology and progression. Preliminary studies using AA subjects who have undergone CSF analysis (n=36) show that AA MCI subjects are more likely to have normal CSF AD biomarkers than NHW MCI subjects, yet at the same time greater hippocampal atrophy on MRI. We hypothesize that endothelial dysfunction is an alternate mechanism which independently contributes to cognitive impairment in AA subjects with sub-threshold AD pathology in a race-independent fashion, and endothelia dysfunction further enhances the neurotoxicity of AD- associated brain changes in a race-dependent fashion. We propose to build on our success in recruiting AA volunteers into memory and aging studies at the Emory's Registry for Remembrance to recruit a cross- sectional cohort of 75 AA subjects along with 75 NHW subjects with normal cognition, MCI, or mild AD. We will test our hypothesis through two aims. In Aim 1, we will determine whether endothelial dysfunctions independently contribute to cognitive decline in AA and NHW subjects by measuring cerebrospinal fluid (CSF) levels of AD, endothelial, and inflammatory markers. Each subject will also undergo MRI analysis for total area of white matter hyperintensities as an imaging marker of endothelial dysfunction. Based on our hypothesis, we predict that AA MCI/AD subjects are more likely than NHW MCI subjects to have normal CSF AD biomarkers, abnormal CSF endothelial markers, and greater number and area of white matter hyperintensities on MRI. In Aim 2, we will determine if an endothelial marker - intercellular adhesion molecule 1 or ICAM-1 - gene variant unique to AA enhances AD neurotoxicity to explain the greater hippocampal atrophy among AA MCI subjects. The Lys56Met ICAM1 gene variant associated with low ICAM-1 levels is uniquely found in 16-20% of AA, and these subjects may have impaired downstream activation of neprilysin, an Ab-degrading enzyme. If our hypothesis is true, AA subjects with the Lys56Met gene variant will be more likely to have hippocampal atrophy, temporal-parietal cerebral hypoperfusion, and cerebral amyloid deposition than AA subjects and NHW subjects without the gene variant. This may occur in the setting of CSF Ab42 pseudo-normalization if low neprilysin levels lead to increased Ab42 levels.